Description
The confinement performance of H-mode plasmas depends on the pressure profile in the pedestal region. EPED1 model is a widely used pedestal model, which determines the pedestal height and width based on MHD stabilities and turbulence without calculating energy fluxes. The pressure in the pedestal region with type-I edge localized modes (ELMs) gradually increases in the transport timescale, and then it suddenly decreases when it reaches the upperlimit constrained by MHD instabilities and turbulence. Therefore, transport should be solved while considering the MHD and turbulence stabilities that trigger ELMs and the change in the magnetic equilibrium for the time-dependent pedestal simulations.
In this study, we propose an empirical transport model for the edge region. It was found that the ELM frequency depends on the heating power, and the loss of the stored energy due to ELMs is almost constant in the JT-60U experiment [1]. Based on the observations, we assume that the ELM frequency and the loss of the stored energy can be estimated by a given heating power, and predict the heat diffusivity in the inter-ELM phase and the amount of the temperature reduction due to ELMs to match the observations. Here, the density profile is fixed. In the transport simulation with the empirical transport model, the MHD stability analysis is required at every time step to trigger the ELM. For the rapid MHD stability analysis, we have constructed a neural-network (NN) based surrogate model of the ideal MHD stability code MARG2D [2]. The empirical transport model and MARG2D-NN are introduced into the transport code TRESS and are used to perform integrated simulations of H-mode plasmas with ELMs.
In the integrated simulations, the pedestal width has been fixed so far. Also, the magnetic equilibrium is not solved, although the change in the bootstrap current is considered. A turbulence stability model for the pedestal width and a magnetic equilibrium solver will be developed, and transport, MHD and turbulence stability, and magnetic equilibrium coupling simulations will be presented.
[1] H. Urano et al., Nucl. Fusion 53 083003 (2013).
[2] N. Aiba et al., Comput. Phys. Commun. 175 269 (2006).